Commit Graph
100 Commits
Author SHA1 Message Date
Alexandre Teixeira 576abb012d feat(browser): deterministic private_browser lifecycle (Wave 5A)
Own each agent-browser session as a browser tree: the daemon's POSIX
session, its runtime files and its Chrome profile. Timeouts, launch
failures, bootstrap recovery, cancellation and shutdown clean that tree
and verify nothing survives, instead of killing only the daemon and
orphaning Chrome. Per-call cleanup no longer sweeps every Chrome under
the runtime TMPDIR.

Sessionless calls get an ephemeral browser closed before returning.
Actions on one session are serialized. Recovery is bounded by one
deadline with at most one retry for local HTML open, and the retry flag
is no longer model-visible. Observations after a failed navigation are
marked stale. read URL navigates and extracts in one batch because
agent-browser has no read command. Results carry a browser_lifecycle
receipt with stages, timings, ownership and cleanup evidence.

Playwright MCP tool calls are bounded by
ODYSSEUS_BROWSER_MCP_CALL_TIMEOUT_S and are not retried. research_navigator
now passes timeout_ms.
2026-10-01 20:59:27 +01:00
Alexandre Teixeira cb5b81022b Merge pull request #50 from pewdiepie-archdaemon/feature/runtime-request-authority
feat(runtime): enforce server request authority
2026-10-01 19:50:38 +01:00
Alexandre Teixeira 9d0257134f feat(runtime): enforce server request authority 2026-10-01 18:35:41 +01:00
Alexandre Teixeira 05c6cfcfa1 Merge pull request #43 from pewdiepie-archdaemon/feature/agent-runtime-wave-1-1
feat(runtime): complete Agent Runtime Wave 1.1
2026-10-01 15:22:06 +01:00
Alexandre Teixeira d49071bbec fix: close Wave 1.1 completion-gate audit findings
- Headless consumers (task scheduler, background follow-up) now treat a
  completion-gate final_response as the authoritative answer instead of
  collecting deltas only. A gated replacement no longer leaves scheduled
  output empty, which used to trigger an extra, ungated grace-summary
  model call.
- The scheduler closes the agent stream with contextlib.aclosing, so the
  approval-pause break unwinds the gate's journal and teacher-takeover
  context in its own task. Chained runs no longer inherit a stale
  parent_run_id, and later finalization no longer raises ContextVar
  reset errors.
- On provider error, the completion gate applies the live answer's
  statement filter to persisted round_texts. Diagnostics and the failure
  note survive; claims rejected by the gate cannot reappear on reload.
2026-10-01 14:49:32 +01:00
Alexandre Teixeira f4793696f4 merge: reconcile Wave 1.1 with post-PR40 lab
Merge canonical lab 9557b8d5909eb4a885c3bf49e19a65dd904f8c1d exactly once.
Retain invocation journal ownership and lineage, provider terminal ordering,
teacher handoff, framed DONE handling, and canonical authority/Ajax routing.

Combine dynamic dispatch receipts with lab policy forwarding. Adapt native
shell/patch evidence, explicit TUI verifiers, and artifact recovery presentation.
Refresh generated configuration source links and strengthen adapter regressions.

Validation: focused 2118 passed; Wave 1.1 script 2291 passed; broad runtime
5649 passed; full pytest 11581 passed, 53 skipped, 2 xfailed, 6 subtests passed.
Compileall 1689 Python files; syntax 279 JS and 82 MJS files; diff and
conflict-marker checks passed.
2026-10-01 09:09:55 +01:00
Alexandre Teixeira e5e23e640d Merge pull request #40 from pewdiepie-archdaemon/review/harness-latest-20260917
Review preview harness, editor, email and task improvements
2026-10-01 06:59:19 +01:00
Alexandre Teixeira 5dfe1c353c test: stabilize final PR 40 validation gates 2026-10-01 06:39:22 +01:00
Alexandre Teixeira f74a262f73 merge: reconcile PR 40 with current lab
Integrate lab fff55a78 into PR #40 (cc25d5ba). Lab's modular email
backend/frontend, modular settings, split stylesheets (static/style.css
stays deleted), procfs compatibility, and request-scoped TurnContract
authority win; PR #40's routing classifiers, editor/email/task features,
and style.css changes are ported into lab's module and stylesheet homes.

Integration fixes:
- settings/api.js imports ui.js under its canonical versioned URL
- browser observations keep legacy CAPTCHA/access-block evidence
- artifact turns do not re-trigger broad-web research recovery
- env reference documents PR test-tool variables; page regenerated

PR #40 defects surfaced by lab gates and fixed here:
- web_fetch generic schema drops top-level anyOf (OpenAI contract);
  the compact preview contract still requires url or urls
- get_weather registered as a brokered network read
- new lazy editor modules precached for offline use
- SearXNG pin mirrored into GPU standalone compose files
- image model picker again skips offline endpoints

Tests updated where PR #40 changed behaviour on purpose, and PR tests
moved onto lab's document_source helpers.
2026-10-01 05:03:58 +01:00
Alexandre Teixeira f93767d5c7 Merge pull request #17 from o3LL/fix/procfs-pid-file-liveness
fix(browser): stop treating a missing cmdline as proof the daemon exited
2026-10-01 03:10:43 +01:00
Alexandre Teixeira d63932f1ad Merge lab into fix/procfs-pid-file-liveness 2026-10-01 03:10:10 +01:00
Alexandre Teixeira a462e01483 Merge pull request #39 from o3LL/feat/admin-build-provenance
feat(settings): show the running build's version and commit in the admin panel
2026-10-01 02:55:54 +01:00
Alexandre Teixeira 6901c56192 Merge lab into feat/admin-build-provenance 2026-10-01 02:55:27 +01:00
Alexandre Teixeira f817ab6e09 Merge pull request #35 from o3LL/docs/env-configuration-reference
docs: generate the ODYSSEUS_* configuration reference from the source
2026-10-01 02:53:07 +01:00
Alexandre Teixeira c6f690a27b Merge lab into docs/env-configuration-reference 2026-10-01 02:52:37 +01:00
Alexandre Teixeira 762fb891ce Merge pull request #29 from o3LL/test/95-release-smoke-suite
test(smoke): add a release smoke suite over every advertised feature area
2026-10-01 02:46:16 +01:00
Alexandre Teixeira bdccb1ddb1 Merge lab into test/95-release-smoke-suite 2026-10-01 02:45:49 +01:00
Alexandre Teixeira 95416cdbfb Merge pull request #27 from o3LL/audit/ref-parity
chore(tools): add a read-only ref-to-ref parity audit
2026-10-01 02:45:38 +01:00
Alexandre Teixeira ef0d96a3ae Merge lab into audit/ref-parity 2026-10-01 02:45:10 +01:00
Alexandre Teixeira fe7297547c Merge pull request #38 from o3LL/test/runtime-behavior-regressions
test(runtime): pin negative capability wording, test-only
2026-10-01 02:42:50 +01:00
Alexandre Teixeira 8c7e3a9411 Merge lab into test/runtime-behavior-regressions 2026-10-01 02:42:22 +01:00
Alexandre Teixeira 1dca85cb07 Merge pull request #37 from o3LL/fix/token-cache-atomic-swap
fix(auth): swap the API token cache atomically instead of clearing it
2026-10-01 02:42:09 +01:00
Alexandre Teixeira 45e1e2e7d9 Merge lab into fix/token-cache-atomic-swap 2026-10-01 02:41:41 +01:00
Alexandre Teixeira bf76c9c608 Merge pull request #36 from o3LL/fix/6174-singleflight-cancel
fix(tasks): clean up the singleflight cache on cancellation
2026-10-01 02:28:18 +01:00
Alexandre Teixeira b0bc0b8c40 Merge lab into fix/6174-singleflight-cancel 2026-10-01 02:27:49 +01:00
Alexandre Teixeira 236baea232 Merge pull request #33 from o3LL/refactor/email-library-package
refactor(email): decompose emailLibrary.js into a package behind a re-export wrapper
2026-10-01 02:23:56 +01:00
Alexandre Teixeira 799adbbde6 Merge lab into refactor/email-library-package 2026-10-01 02:22:23 +01:00
Alexandre Teixeira eef390b9ae Merge pull request #22 from o3LL/refactor/routes-email-subpackage
refactor(routes): move the email modules into routes/email/
2026-10-01 02:17:29 +01:00
Alexandre Teixeira 9477616e05 Merge lab into refactor/routes-email-subpackage 2026-10-01 02:15:36 +01:00
Alexandre Teixeira aedec7d005 fix(runtime): isolate nested invocation ownership 2026-10-01 02:11:53 +01:00
Alexandre Teixeira b35010c3d1 Merge pull request #24 from o3LL/refactor/settings-shell-modules
refactor(settings): move the shell out of settings.js into settings/
2026-10-01 01:42:18 +01:00
Alexandre Teixeira 33387d4a4c Merge lab into refactor/settings-shell-modules 2026-10-01 01:41:27 +01:00
Alexandre Teixeira 4c122de880 fix(runtime): scope completion claims to execution obligations 2026-10-01 01:36:18 +01:00
Alexandre Teixeira 290bb0d61d Merge pull request #23 from o3LL/refactor/settings-panel-modules
refactor(settings): extract four panels into settings/ modules
2026-10-01 01:08:58 +01:00
Alexandre Teixeira 9ee9205bfc merge: sync lab after stylesheet split 2026-10-01 01:08:15 +01:00
Alexandre Teixeira 86212bfe99 Merge lab into refactor/settings-panel-modules 2026-10-01 01:03:49 +01:00
Alexandre Teixeira 0f23439a31 Merge pull request #20 from o3LL/refactor/split-style-css
refactor(css): split style.css into ordered fragments
2026-10-01 00:55:35 +01:00
Alexandre Teixeira 23fdc26079 Merge lab into refactor/split-style-css 2026-10-01 00:48:58 +01:00
Alexandre Teixeira 466a6b323a fix(runtime): preserve provider error terminal ordering 2026-10-01 00:33:46 +01:00
Alexandre Teixeira fe80eb795e merge: reconcile reviewed lab baseline for runtime wave 1.1 2026-10-01 00:25:42 +01:00
Alexandre Teixeira 0a88b3c322 Merge pull request #31 from o3LL/test/document-module-set-helper
test(document): read the editor through a module-set helper before it is split
2026-09-30 22:59:49 +01:00
Alexandre Teixeira 40dbe17a95 Merge lab into test/document-module-set-helper
Resolve the overlap with #19 by preserving the whole-cascade stylesheet
helpers alongside #31's document-source/module-set helpers.

Maintainer validation:
- document/module contract and composition guards: 15 passed
- all 54 touched Python test modules: 366 passed
- py_compile: clean
- git diff --check: clean
2026-09-30 22:56:56 +01:00
Alexandre Teixeira 741f5dff2b Merge pull request #19 from o3LL/refactor/tests-read-the-whole-cascade
refactor(tests): read the whole cascade instead of style.css alone
2026-09-30 19:27:58 +01:00
Alexandre Teixeira 6ee51fee72 Merge pull request #18 from o3LL/fix/duplicate-keyframes
fix(css): collapse duplicate @keyframes names to the definition that wins
2026-09-30 19:18:13 +01:00
Alexandre Teixeira a969a68016 Merge pull request #34 from o3LL/test/macos-failures-and-stub-leak-guard
test: fix environment-dependent failures and guard module-stub leaks
2026-09-30 19:16:43 +01:00
Alexandre Teixeira 56484df737 test(media): detect ffmpeg encoders by codec alias 2026-09-30 19:16:29 +01:00
Alexandre Teixeira eaddc03729 Merge pull request #26 from o3LL/fix/tmp-realpath-test-bugs
fix(tests): resolve temp paths consistently on macOS
2026-09-30 18:21:04 +01:00
Alexandre Teixeira bfae249120 Merge pull request #28 from o3LL/fix/cookbook-stop-procfs-guard
fix(cookbook): skip the pid sweep when the host has no procfs
2026-09-30 18:07:07 +01:00
Alexandre Teixeira 054df080af Merge pull request #32 from o3LL/fix/6215-mcp-args-validation
fix(mcp): reject malformed Args on Add MCP Server instead of defaulting to []
2026-09-30 17:18:51 +01:00
Alexandre Teixeira 11ecf46abb Merge pull request #30 from o3LL/fix/6228-tailscale-empty-lookup-cache
fix(discovery): cache a successful but empty Tailscale lookup
2026-09-30 17:05:16 +01:00
Alexandre Teixeira 3748a621d1 Merge current lab into agent runtime decomposition 2026-09-29 19:15:54 +01:00
Alexandre Teixeira 6105702901 Merge pull request #16 from o3LL/lane/decomposition-and-test-isolation
refactor(lane): decompose style.css, pin it with a computed-style harness, and isolate worktree runs
2026-09-29 14:54:36 +01:00
Alexandre Teixeira 65c35c8211 Merge pull request #15 from o3LL/fix/test-static-port-isolation
test(harness): bind the static test server to an ephemeral port
2026-09-29 14:54:18 +01:00
Alexandre Teixeira 956671a731 Merge pull request #14 from o3LL/fix/proc-guard-private-browser
fix(browser): skip the Chrome sweep when the host has no procfs
2026-09-29 14:54:00 +01:00
Alexandre Teixeira aafd8a98eb Merge pull request #13 from o3LL/v1/css-split-3
refactor(css): move cookbook, research, memory and settings styles to their own file
2026-09-29 14:53:44 +01:00
Alexandre Teixeira cea8ed297e fix(runtime): reject unobserved verification and preserve safe reasoning 2026-09-26 14:24:29 +01:00
Alexandre Teixeira b241bb3a7b fix(runtime): close completion stream bypass and preserve explanations 2026-09-26 14:13:40 +01:00
Alexandre Teixeira eaa5668baa feat(runtime): record execution evidence and gate completion 2026-09-26 13:48:12 +01:00
Alexandre Teixeira ea847e6c0b test(runtime): establish isolated validation and comparison gates 2026-09-26 13:13:07 +01:00
Alexandre Teixeira 0e07d9a675 Merge pull request #10 from o3LL/fix/review-20260923-bound-outbound-and-draft-limits
fix(security): harden scholarly lookups and editor draft limits
2026-09-24 12:22:04 +01:00
Alexandre Teixeira 11ad4e2739 fix(url-safety): resolve NAT64 well-known prefix to IPv4 target
The R09 scholarly hardening routes every hop through check_outbound_url with
block_private=True. On a DNS64/NAT64 network, an IPv4-only host can resolve
through the RFC 6052 Well-Known Prefix. For example, export.arxiv.org resolved
to 64:ff9b::924b:5b2a in the reproduced environment.

CPython classifies that outer IPv6 prefix as reserved, so the URL guard rejected
the request before examining the effective IPv4 destination.

Decode addresses in exactly 64:ff9b::/96 to their embedded IPv4 destination and
evaluate that destination under the strict outbound policy.

The translated target is always checked with private-address blocking enabled.
This prevents the NAT64 prefix from becoming a path to loopback, private,
shared/CGNAT, link-local, multicast, unspecified, or other non-global IPv4
space.

Network-specific translation prefixes are not decoded. In particular,
64:ff9b:1::/48 remains subject to the existing IPv6 policy.

Coverage includes:

- public IPv4 destinations embedded through the RFC 6052 prefix
- loopback, link-local, private, CGNAT, multicast, unspecified, benchmark, and
  TEST-NET rejection
- network-specific NAT64 prefixes remaining undecoded
- deterministic scholarly provider tests without live DNS dependence
- redirect query parameter isolation
- relative redirect resolution
- HTTP error fallback behavior
- editor draft GET and DELETE behavior remaining unaffected

R09, R11, and R01 were independently audited and otherwise left unchanged.
2026-09-23 15:32:04 +01:00
Alexandre Teixeira 0784cd2fed fix(review): harden scholarly redirects, editor draft limits, and threat model
- Enforce outbound URL policy on all hops via bounded manual redirects in scholarly lookups
- Guard declared Content-Length in EditorDraftRoute before request body parsing
- Keep scholarly lookup timeouts and budgets as internal constants rather than surface env vars
- Narrow TUI threat-model description around demonstrable host shell bridge behavior
- Add behavioral regression tests for redirect security and pre-parsing body size guards
2026-09-23 12:53:06 +01:00
Alexandre Teixeira c90a81dbdc Merge commit '14afd3afb6274ff986733ddff992001e19c80e29' into fix/pr10-merge-ready 2026-09-23 12:29:27 +01:00
Alexandre Teixeira f577934777 fix(ci): restore green lab baseline before runtime W1 (#9)
Validated locally and on GitHub Actions before integration into lab.
2026-09-23 10:16:54 +01:00
Alexandre Teixeira 771d48f147 ci: install FFmpeg for media integration tests 2026-09-23 01:08:20 +01:00
Alexandre Teixeira 3cabbb9bca test(agent): restore exact turn-policy regressions 2026-09-23 00:14:30 +01:00
Alexandre Teixeira c0b71a5cef fix(tools): constrain Python environment namespace mount 2026-09-23 00:10:50 +01:00
Alexandre Teixeira 0012baa17c ci(trivy): free build cache before image scan 2026-09-22 23:27:29 +01:00
Alexandre Teixeira c0a1cecc00 test(baseline): align UI and model profile expectations 2026-09-22 23:27:17 +01:00
Alexandre Teixeira 9bc1cdcab0 test(harness): isolate imports and stabilize browser and DNS fixtures 2026-09-22 23:27:08 +01:00
Alexandre Teixeira 4520b4f6f0 fix(ui): restore rich text controls and accessible research map 2026-09-22 23:26:43 +01:00
Alexandre Teixeira b9cccb93ac fix(tools): expose active Python environment in workspace namespace 2026-09-22 23:26:28 +01:00
Alexandre Teixeira 79a55fac38 fix(agent): preserve focused turn contracts and verified completion 2026-09-22 23:26:16 +01:00
Alexandre Teixeira 128102ef26 test(runtime): isolate tool execution module state 2026-09-22 18:17:42 +01:00
Alexandre Teixeira 5568d7d631 fix(sw): complete editor panel precache 2026-09-22 18:09:04 +01:00
Alexandre Teixeira 83088f8811 ci: provision browser dependencies for pytest 2026-09-22 18:08:16 +01:00
Alexandre Teixeira ebcc6c524f test(database): restore model endpoint isolation 2026-09-22 18:07:09 +01:00
Alexandre Teixeira dbba4978ee test(harness): support cache-busted frontend module imports 2026-09-22 18:06:18 +01:00
Alexandre Teixeira 7e798fe925 Merge pull request #6 from pewdiepie-archdaemon/pr/alteixeira20/subscription-provider-ux
feat(provider): improve subscription UX and lazy discovery
2026-09-22 14:02:55 +01:00
Alexandre Teixeira 8bb48780a1 feat(provider): add lazy Featherless model discovery 2026-09-22 13:12:19 +01:00
Alexandre Teixeira 2b2bf0fb90 fix(ui): refine provider endpoint controls 2026-09-22 13:12:19 +01:00
Alexandre Teixeira 46c8451a29 fix(ui): bust cache for collapsible subscription usage 2026-09-22 13:12:19 +01:00
Alexandre Teixeira 411559913c feat(ui): refine subscription model and usage controls 2026-09-22 13:12:19 +01:00
Alexandre Teixeira ed7ccfd584 feat(provider): support multiple ChatGPT subscriptions with usage 2026-09-22 13:12:19 +01:00
Alexandre Teixeira 45330097b8 Merge pull request #7 from pewdiepie-archdaemon/fix/maintainer-harness-ci-reproducibility-v1
fix(ci): make maintainer harness and lab workflow portable
2026-09-22 12:44:20 +01:00
Alexandre Teixeira dd13f53507 ci: support maintainer lab pull requests 2026-09-22 12:39:14 +01:00
Alexandre Teixeira 822ceaaac4 fix(ci): make maintainer harness contract self-contained and portable 2026-09-22 11:31:45 +01:00
Alexandre Teixeira cee319050c refactor(settings): add registry-backed navigation and finder (#6040)
* refactor(settings): add modular shell primitives

* refactor(settings): wire modular shell

* test(settings): exercise real coordinator ESM boundary

* refactor(settings): add registry-backed settings finder

* fix(settings): harden registry navigation behavior
2026-08-16 02:48:19 +01:00
Alexandre TeixeiraandRaresKeY 9c71948376 fix(companion): honor configured pairing address (#6060)
* fix(companion): honor configured pairing origin

* fix(companion): keep configured pairing on v1 LAN contract

* fix(companion): reject numeric pairing hosts

---------

Co-authored-by: RaresKeY <158580472+RaresKeY@users.noreply.github.com>
2026-08-15 21:11:10 +02:00
c4369305f0 refactor(model-routing): centralize explicit foreground fallback policy (#6020)
* refactor(model-routing): centralize explicit foreground fallback policy

Make foreground fallback an explicit per-user, availability-only policy shared by streaming Chat, non-stream Chat, and Agent runs.

Preserve strict defaults, owner/model and credential boundaries, pinned Agent routes, and truthful per-round provenance/accounting. Carry provider-reported model identifiers through native streaming adapters, non-stream responses, and caches, and keep legacy default_model_fallbacks as tombstoned raw storage that generic settings APIs and agent tools cannot expose or mutate.

* fix(agent-loop): restore rebase-dropped qwen routing, workspace prompt, and temperature clamp

* fix(model-routing): thread selected endpoint identity, fix cost classification and fallback eligibility

* fix(chat): restore stream helpers and harden run stop lifecycle

* fix(model-routing): let numeric provider codes win over symbolic rate-limit statuses

* fix(agent-loop): apply qwen temperature and notes-tool clamps per fallback candidate

* fix(chat): honor queued stop across resend and reload canonical terminal on EOF

* fix(chat): track stop queue and cleanup ownership by per-send generation

* fix(agent-loop): preserve requested temperature for non-qwen fallback candidates

* fix(chat): reserve send ownership before any await and scope stop to the current send

* fix(chat): clear the previous run identity at send reservation

---------

Co-authored-by: RaresKeY <158580472+RaresKeY@users.noreply.github.com>
Co-authored-by: StressTestor <212606152+StressTestor@users.noreply.github.com>
2026-08-14 08:10:30 +01:00
Alexandre Teixeira e466b02f1a fix(security): make research path lookup CodeQL-friendly (#5129)
* fix(security): make research path lookup CodeQL-friendly

* fix(security): avoid duplicate research path scans

* fix(research): preserve active completed spinoff query
2026-07-04 20:17:45 +02:00
Alexandre Teixeira 5b4ed45e69 Merge pull request #5195 from ashvinctrl/fix/send-to-session-null-owner
fix(security): scope send_to_session to an exact session owner
2026-07-04 16:47:10 +01:00
Alexandre Teixeira 972c2d6c62 ci: add focused test guidance signal (#4982)
* ci: add focused test guidance signal

* ci: diff focused guidance from merge base
2026-07-03 21:17:28 +02:00
Alexandre Teixeira 667cc4e2fa test: split service health tests (#4972)
* test: split service health tests

* test(service-health): preserve focus selector
2026-07-03 20:50:49 +02:00
Alexandre Teixeira a24e749404 fix(security): confine research file paths (#4986) 2026-07-02 11:58:35 +01:00
Alexandre Teixeira e86ccae0db fix(docker): make host Docker socket opt-in (#4902)
* fix(docker): make host socket compose opt-in

* fix(cookbook): gate container Docker access

* fix(docker): gate socket group setup on opt-in

* fix(cookbook): gate generated docker exec serve commands

* fix(cookbook): narrow generated docker exec forms
2026-06-30 19:54:51 +02:00
Alexandre Teixeira 4c7db7c216 fix(security): harden gallery endpoint URL checks (#4981)
Replace substring OpenAI endpoint detection with exact parsed-host matching.

Route gallery image endpoint construction through a constant path allowlist.

Remove client-visible exception and upstream response body leaks from gallery image flows while preserving diagnostics in server logs.

Add focused regression tests for OpenAI host matching, checked endpoint joining, harmonize SSRF hardening, and sanitized client errors.
2026-06-30 19:16:34 +02:00
Alexandre Teixeira c8e106c629 test: split provider endpoint tests (#4961) 2026-06-28 19:05:38 +02:00
Alexandre Teixeira 31705015a3 test: localize calendar recurrence helper import (#4944)
* test: localize calendar recurrence helper import

* test: share calendar route import helper
2026-06-28 19:04:15 +02:00
Alexandre Teixeira b2881ea0d1 test: split endpoint resolver tests (#4957) 2026-06-28 00:49:43 +02:00